Chiliparrot has joined #nix-darwin
Chiliparrot has quit [Ping timeout: 260 seconds]
stephank5 has joined #nix-darwin
stephank has quit [Ping timeout: 256 seconds]
stephank5 is now known as stephank
Chiliparrot has joined #nix-darwin
<johnw> ok, not just me at least
<johnw> i tried to git bisect it, but a LOT of commits between my "last good" commit and the current head of unstable had a variety of problems
<johnw> so it's not clear when things went off the rails
<LnL> moin
<LnL> I haven't updated since the python 3.8 update
<johnw> ah, I saw that at some points, that had an "infinite recursion encountered" probleb for me
<johnw> but that appears to have later been resolved
<johnw> moin LnL!
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<stephank> I noticed some installer talk, and I have opinions. 🙃 I built this a little while ago, and I'm wondering if I should make PRs to get this (or parts) into nix-darwin proper:
<stephank> The install process has some confusing flags/questions about apfs, multi-user, or /etc stuff. A nix-darwin specific example: 'Would you like to manage <darwin> with nix-channel?' I had no idea what that meant as a new user.
<stephank> What I tried to do with the alternate installer is have just one recommended procedure / installer result, and do as many checks as possible up front so we can tell the user exactly what's about to happen.
<stephank> But also, I honestly think it'd be awesome if nix-darwin was the official way to get Nix on mac. It means all the system configuration Nix needs (launchd plist, users, etc.) become managed, and it can upgrade itself and stuff.
Chiliparrot has joined #nix-darwin
<LnL> we talked about that a while ago, very cool stuff
<stephank> Ah, I see the logs! Do I understand correctly, your idea is to have a 'nix-installer' that actually borrows a small subset of nixos/nix-darwin to do the work? (Not rely on nix-darwin at all, but nix-darwin would coexist somehow.)
__monty__ has joined #nix-darwin
<LnL> not sure if they have to coexist, but if it's compatible then it's easy to go from one to the other
<LnL> in it's current form, especially for the daemon there's nothing that owns the installation
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
Chiliparrot has joined #nix-darwin
__monty__ has quit [Quit: leaving]
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
eraserhd3 has joined #nix-darwin
eraserhd2 has quit [Ping timeout: 256 seconds]
Chiliparrot has joined #nix-darwin
<johnw> LnL: does this build for you: python2Packages.libxml
<johnw> it's missing libintl
<johnw> but when I try to add it, I get:
<LnL> nope
<johnw> output '/nix/store/7czx189wk20ss9yd1dw62s51r16fq83c-bootstrap-stage2-stdenv-darwin' is not allowed to refer to the following paths:
<johnw> /nix/store/f73jl5d5x4csv6n5p19643d35xly9hgm-gettext-0.20.1
<johnw> /nix/store/yqqpnmzvfmki829n705yw7b2x2vh3fyv-libiconv-osx-10.12.6
<johnw> this worked in 9d0c3ffe6783d59b427d018e8341e0084737fde9
<LnL> ah yeah, it's part of the stdenv closure
<LnL> what if you just override it?
<johnw> how? override what?
<LnL> .overrideAttrs instead of changing libxml2 for the entire nixpkgs tree
<johnw> but where would I do it?
<johnw> there's also a global attribute called libxml2Python
<LnL> I usually just try it out in the repl first
<johnw> yes, but _what_ would you override?
<johnw> libxml2 works, python2Packages.libxml2 doesn't
<johnw> so do I change the definition of python2Packages?
<johnw> do I also fix libxml2Python?
<LnL> not sure, but I think that's just an alias for the python output
<johnw> ah, it's libxml2Python that gtksourceview is trying to pull in
<LnL> hm, wonder why that uses python2
<johnw> should I just push this to master?
<LnL> nix-repl> :b (libxml2.override { python = python2; }).overrideAttrs (drv: { buildInputs = [ gettext ]; })
<johnw> well, libxml2 works fine
<johnw> it's python2Packages.libxml2 that doesn't
<LnL> yeah, that libmlx2Python thing is kind of messy
<LnL> and not correct I think
<johnw> i'll just push to master
<LnL> I would expect this to be the case, which is nonsensical
<LnL> line2: pyhthon2Packages.libxml2.buildInputs -> libxml2 -> python2
<LnL> line13 libxml2 -> python3
<johnw> grr.. master has already changed this in a way that makes it harder to verride
<LnL> also pkgs.libxml2 itself, not the python package fails with python2
<johnw> I guess that deserves a bug report
<LnL> so changing the expression there to include it based on python.pythonVersion might make sense
<johnw> what's the attribute?
<LnL> not sure it's exposed directly
<johnw> which expression where?
<LnL> pkgs.libxml2.override { python = python2; }
<johnw> ah
<LnL> which is the one that's used within the python2Packages context
<johnw> I see
<johnw> well, fixing libxml2Python fixes me
<johnw> so not sure I'll worry about it
<LnL> there are already python version conditionals in libxml2 so that's probably the easiest place to fix
<johnw> ok, that works better
__monty__ has joined #nix-darwin
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<johnw> will push this fix once my build finishes
<johnw> wonder how to fix this zzip problem with texlive now...
<johnw> apparently in this interim the libzzip expression changed pretty significantly
<wildsebastian> johnw It seems there is a PR open for the zzip build error
<{^_^}> #91260 (by veprbl, 2 days ago, open): zziplib: convert back to autotools build
philr_ has quit [Ping timeout: 264 seconds]
evelyn has quit [Ping timeout: 265 seconds]
evelyn has joined #nix-darwin
__monty__ has quit [Quit: leaving]
mbrgm_ has joined #nix-darwin
mbrgm has quit [Ping timeout: 240 seconds]
mbrgm_ is now known as mbrgm
philr_ has joined #nix-darwin